G. H. STEMPEL, JR.,
1158
AND
I n one instance the alkali-insoluble fraction was isolated by drying and evaporating the ether solution. From the residue was obopined a 2,4dinitrophenylhy&arnne meltIts equal mixture with authentic benzoing at 238-!239 phenone 2,4-dinitrophenylhydrazone (m. p. 238-239 ”) melted at 238-239’. Hvdroceno~sof Catechol Diohenvlmethvkne Ethera -Pdld%m ohide was prepared k d h g to ;he directions of Starr and Hixon.” Ten mg. of this catalyst took up 2.5 cc. of hydrogen. The activity of the catalyst with respect to the splitting of benzyl ethers was established by the hydrogenolysis of benzyl tetra-acetyl-8-glucoside ([aID -52.3’) kindly furnished for the purpose by Dr. N. K. Richtmyer. This compound was cleaved quantitatively under the conditions already described by Dr. Richtmyer and in duplication of his results.6 To 1.0 g. of the catechol diphenylmethylene ether in 100 cc. of absolute alcohol was added 0.5 g. of palladium oxide. This mixture took up the theoretical quantity of hydrogen in ten to twelve hours. After completion of the reduction the alcohol was evaporated; the residue was dissolved in ether and extracted as described above. The sublimed products possessed the melting points of the pure catechols and these were not depressed by mixture with the corresponding authentic catechol. The yields were 8585% of the theoretical. The alkali-insoluble fraction in the ether solution was a low melting solid, presumably diphenylmethane, which was not further investigated.
.
(14) Starr and Hixon, “Organic Syntheses,” A. H. Blatt, Editor, John Wiley and Sons, N e w York, N. Y.,1943, Coll. Vol. 11, p. 666.
[CoNTRIBUTXON FROM THE
GERSONS. SCHAFFEL
Vol. 66
Action of Alkali on Catechol Diphenylmethylene Ethers. --One gram of the cyclic ether was added t o 25 cc. of a %.% solution of potassium hydroxide in methanol and the mixture, protected from atmospheric carbon dioxide, waz refluxed for seventy-two hours. It was then cooled at 0 for forty-eight hours; the crystals which formed were filtered off and washed with ice-cold methanol. The product was identical with the starting material and was recovered in 9045% yields. Action of the Grignard Reagent on Catechol Diphenylmethylene Ethers.-The Grignard reagent was formed from 1 g. of 1-bromobutane, b. p. 1OO-10lo, and 0.2 g. of magnesium turnings in 50 cc. of dry ether. To this reagent. protected by dry nitrogen, was added 1 g. of the catechol diphenylmethylene ether; the mixture was refluxed for forty-eight hours. At the end of this period, the solution was thrown on ice and exactly neutralized with 1 N hydrochloric acid. The ether layer was dried and evaporated and the residue was recrystallized from absolute alcohol. The original cyclic ether was recovered in 90-9595 yield.
summary The preparation and properties of the diphenylmethylene ethers of catechol, 3-n-propylcatechol, and 4-t-butylcatechol, have been studied. Although these ethers are resistant to alkali and to the Grignaxd reagent, they are cleaved by dilute mineral acid and by catalytic hydrogenation. BETHESDA, MARYLAND
RECEIVED APRIL 17, 1944
DEPARTMENT OF CHEMISTRY, CARNEGIE INSTITUTB OF TECHNOLOGY]
A Comparative Study of the Kinetics and Mechanisms of Formation of the Phenylhydrazone, Semicarbazone and Oxime of d-CarvoneL2 BY G. H. STEMPEL, JR., AND GERSON S. SCHAFFEL The assumption that the reactions of a ketone does not bring about racemization6 as it does in with the carbonyl reagents, such as phenylhydra- the case of I-menthone. In order to compare these reactions, the followzine, semicarbaizide and hydroxylamine, proceed by identical mechanisms has heretofore been ing four characteristics were studied: (1) effect generally accepted without direct experimental of added neutral salts upon the velocity constants; evidence.*.‘ The object of this investigation was (2) dependence of their rates upon the carvone to determine the correctness of this assurpption concentration; (3) dependence of their rates upon by means of a comparative study of the lunetics carbonyl reagent concentration ; and (4) .deterof these reactions, as well as to obtain data which mination of whether the reactions are general or could lead to a better picture of such a common s p e d c acid catalyzed. mechanism. Experimental The kinetics of these reactions were followed Reqenta-The &one used in these experimentspras by observing the change of rotation during the a n Eastman Kodek Co. product which was further purlfied reactions of the carbonyl reagents with the op- by the method of Wallach6 through the formation of a tically active ketone, d-canrone. One of the “hydrosulfide” with hydrogen sulfide which was recrystallized several times from alcohol and then reconverted to more important considerations which led to the -one by refluxing with aqueous alkali. The regenerchoice of d-carvone is that the optically active ated dcarvone was carefully distilled, after which it excenter of the molecule is not alpha to the carbonyl hibited the following properties: b. p. 230°, corrected group, and consequently enolization of the ketone to 760 mm. pressure; [ P ] = u ~ 73.33’; n% 1.5005 * (1) P r a t e d before the Divioion of Organic Chemistry of the American Chemical Seckty at Detroit in April. 1943. (2) Abstracted from the thesir submitted by Gason S. Schaflel to
the Committee on Graduate IpItnrction at Carncgie Institute of Technology in putial fulEllment of the requirements for the degree of Doctor of Science. (3) C o m t and Bertlett, Tan JOURNAL, 64,2881 (1932). (4) Hommett, “Physicat Organic Chemistry,’‘ McGmr-Hill Bbok C o , lac., New York, N Y , 1940, pp 333-334.
O.OOO3. I n 83% aqueous alcohol, a solvent in which many
of the rate experiments were performed, a 0.4OOO-g. sample
dissolved in sufficient ?% alcohol to make 100 ml. of solu-
tion gave [a 68.53 . The d c a r b a z i d e , phenylhydrazine and hydroxylamine hydrochlorides employed were Eastman Kodak Co. products which were recrystallized several times from (5) B d e t t and Vincent, THISJOVINAL, 66,4992 (1933). (6) Wallach. AIM , 806, 224 (1899).
KINETICSAND MECHANISMS OF ~-CAFWONE WITH CARBONYL REAGENTS
July, 1944
1159
aqueous alcohol and carefully dried over calcium chloride f , a straight line will be obtained whose slope is koa., and sodium hydroxide. The semicarbazone, phenyl-2,303. Figure 1 shows the plot of log (h at)against hydrazone and oxime were prepared, and exhibited th,e time for typical runs with each of the carbonyl reagents. following properties: d-carvone semicarbazone, m. p. 143 , Determinations of the equilibrium constants, K,of these [ ~ y 181.3" l ~ in~ 837, ~ :queous ~ ~ alcohol; d-carvone phenyl- reactions showed that K is of the order of 300 or more and hydrazone, m. p. 110 , [a1905461268.31; in 83% aqueous since k1 = k.b./(l l/Q then kl may be set equal to alcohol; and d-carvone oxime, m. p. 73 , [ a l a 0 ~38.2" ,, in koa. . I n order to obtain the true rate constant, ko, inde: 83yoaqueous alcohol. pendent of the constant carbonyl reagent concentration, The water and alcohol used in making solutions were k1 was divided by the carbonyl reagent concentration. It warmed under reduced pressure to remove dissolved oxy- is this ko that has been recorded in all of the following gen and were stored under an atmosphere of nitrogen in tables. automatic filling burets. All materials used in preparing buffers, etc., were standard c. P. reagents. Apparatus.-The optical measurements were made with a Hilger M-8 polarimeter using the Hg green line (5461 k). On this instrument static readings are repioducible to 0.01 ". The kinetic runs were made in a specially designed Pyrex Polarimeter tube. The tube proper consisted of a 50-cm. Pyrex glass tube to which were sealed flat Pyrex end pieces. At one end of the tube a 100 ml. bulb was attached by means of capillary tubing; a t the other end, a side arm containing a stopcock. The tube was completely enclosed by a water jacket made of 2-inch micarta tubing with end windows placed in brass end caps, and provided with an inlet and outlet permitting the circulation of water thermostatically controlled at 30 * 0.05". Procedure.-When making an experimental run, the solution was prepared in a flask equipped with a gas inlet through which nitrogen was continually bubbled into the Time, minutes. mixture. The carbonyl reagent hydrochloride was weighed into the flask, the requisite quantity of standard potassium Fig. 1.-Typical log (a, at) us. time plots for kinetic hydroxide was added, and the proper amounts of absolute alcohol and distilled water a t 30" were added to bring the runs of d-carvone with: I, phenylhydrazine (A, 0.025; volume to a predetermined value. The d-carvone was B, 0.325); 11, semicarbazide (A, 0.3; B, -0.3); and 111, weighed into a small weighing bottle, which was dropped hydroxylamine (A, 0.0; B, -0.6). into the flask as the timing was begun. The polarimeter Experimental Results tube was thoroughly swept out with nitrogen, the solution was poured into the bulb on the tube while the nitrogen Equilibrium Constants.-Equilibrium conwas still flowing, and the solution was stirred by the gas for a short time. Then the gas flow was discontinued and a stants were obtained from equilibrium rotation slight suction was applied to draw the solution into the readings of typical kinetic runs and were calcutube. The elapsed time from the mixing of the reagents lated from the expression: K = x/(l x)(b until the first reading was taken was from one to two ax), in which a and b are, respectively, the initial minutes. Subsequent readings consisted of the time a t ketone and carbonyl reagent concentrations and which a given rotation was observed. All runs were made a~), where a, is the rotation a t 30' and all concentrations here reported are in moles per x = (ae ao)/(cr, liter. calculated for complete conversibn of ketone to In order to avoid any complicating extraneous effects phenylhydrazone, semicarbazone or oxime. Typiwhich might arise from the use of external buffers, the cal values of K are: for the phenylhydrazine-dpH was maintained constant by the buffering action of the carbonyl reagent and its hydrochloride, according to the carvone reaction: 614, 576 and 572; for the hymethod suggested by Orning and Stempel.' Experimen- droxylamine-d-carvone reaction : 410, 310 and tally, this was accomplished by dissolving a known quantity 380; for the semicarbazide-d-carvone reaction: of the hydrochloride and then adding less than the stoi- 310,330 and 430. chiometric amount of standard base. The concentration Kinetic Salt Effect.-The absence of any effect of the carbonyl reagent, BNH2, was then equal to the concentration of the added inorganic base, and the concentra- upon ko by changing ionic strength is shown in tion of the ion of the hydrochloride, BNHa+, was equal to Table I, in which are shown ko values for a series the difference between the gross initial concentration of BNHa+ and the concentration of the added base. TABLE I By utilization of such a buffer, the concentration of the CONSTANCY OF WITH CHANGING IONIC STRENGTH BNH2 was kept at such a high value that the reaction became pseudo-unirrlolecular. Therefore, the rate equation Concentrations in moles per liter: (A) phenylhydrazine 0.138, phenylhydrazine hydrochloride 0.010, carvone for a first order, reversible reaction expressed in terms of 0.020; (B) semicarbazide 0.054, semicarbazide hydrooptical rotations was used. For the reaction chloride 0.050,carvone 0.022; (C) hydroxylamine 0.088, ki hydroxylamine hydrochloride 0.090, carvone 0.065. RICO BNHs RnC=NB HOH Solvent: A and B, 83% aqueous alcohol; C, 50% aqueous kr alcohol. we have (B)Semicubazide (C) Hydroxylamine (A)r~~cnJrlhydruinc 1 Ionic Ionic k-d = kl k2 ii. - Lnstrength kc X 101 ~trength &D X 10 strength &O X 10 8 &-at 0.010 2.37 0.013 3.12 0.178 2.39 where a. is the final rotation, a0 is the initial rotation, and .020 2.00 .032 3.18 .246 2.36 at is the rotation at any time. NOW,log (a. at) log (a. ao) ( R o b . t/2.3?), ,023 2.13 .041 3.18 .315 2.39 from which it follows that if log (a. at)is plotted agslnst .038 2.13 .067 3.14 (7) Orning and Stempel, J . fig. C h a . . 4, 410 (1939). ,068 2.30
-
+
-
-
-
+
+
+
-
-
-
- -
-
-
G. H.
i iti0
STEMPEI,, JR., riND
of kinetic runs in which all variables except ionic strength were kept constant. The ionic strength was varied by adding calculated amounts of potassium chloride. Dependence of Rate on Ketone Concentration.-In Table I1 are shown the values for ko for the reactions of the three carbonyl reagents with various initial concentrations of d-carvone. The dependence of the rates on the d-capone concentration is shown by the constancy of KO in Table 11.
GERSONs. SCHAFFEL
Vol. 66
make the ratio of the concentration of A- to that of HA the same in all solutions. Since the carbonyl reagents are bases, the following equilibria exist HA @ H + + A -
+
BNH2
P
RSH3+
It is apparent that the effect on the pH by small changes in the concentration of BNH2, which is TABLE I1 used up in the reaction, can be minimized by I-ALLJES OF ko WITH DIFFERENT STARTING CONCENTRATIONS choosing HA so that its acid dissociation constant OF CARVONE is nearly the same as that of BNH3+. A comConcentrations of carbonyl reagent and its hydrochloride parison of K a c i d for the BNHB+ ion with that of both 0.05 mole/liter ; solvent, 8351, aqueous alcohol. the acid of the buffer pair chosen in each case is Phenylhydrazine Hydroxylamine Semicarbazide shown in Table IV. In all runs, the ionic strength Molar Molar Molar concn. of concn. of concn. of was maintained constant by the addition of pocarvone ko X 10 carvone ko X 10 carvone ka X 10 tassium chloride. 0.0062 1.26 0.0065 2.78 0.0203 1'.87 ,0106 ,0140 .0178 .0232
1.24 1.23 1.20
,0158 .0247
2.74 3.72
.0228 .0268
1.88 1.88
1.24
TABLE IV COMPARISON OF DISSOCIATION COhSTANTS OF THE HYDROCHLORIDES OF CARBONYL. REAGENTS AND CERTAIN ORGANIC
ACIDS
Dependence of Rate on Carbonyl Reagent Concentration.-In Table I11 are shown the values of ko for runs with various initial concentrations of carbonyl reagent. These values of ka, found by dividing the observed k by the carbonyl reagent concentration, should remain constant if the rates are proportional to the first power of the carbonyl reagent concentration. That this is approximately true can be seen from the nearly constant values of ko in Table 111. The small drift in ko is attributable to the somewhat increased PH as the concentrations of the basic carbonyl reagents are increased.
Knfor ion,
Carbonyl reagent Phenylhydrazine Semicarhazide Hydroxylamine
BNHI 0 6 X 10-5 2 2 X 10-4 8 X 10-7 +
Acid K. for chosen acid Acetic 1 . 9 x 10Formic 2 . 0 x 10-4 Sodium hydrogen maleate 4 X 10-7
Table V shows the experimental results along with glass electrode readings to show the approximate constancy of the pH in the solutions. It TABLE V
DATASHOWING
ko WITH ACIDITY AT CONSTANT PH Phenylhydrazine Concentration of phenylhydrazine 0.149, of carvone TABLE 111 0.0203 mole per liter; solvent, 80% aqueous alcohol, ionic strength, 0.0719. VALUESOF ko WTH DIFFERENT STARTING CONCEWTRATIONS Acetic acid Acetate ion O F CARBONYL REAGENT concentration concentration ka E. m. f. Concentrations in moles per liter: (A) C&H~NH.NH&~ 0.0128 0.0112 0.0199 0.090 0.05, carvone 0.0203; (B) NH2C0.KH.NHaC1 0.05, ,0256 .0223 .0346 .087 carvone 0.0221; (C) XHzOHCI 0.113, carvone 0.0385; ,0384 .0335 .0468 ,083 solvent, aqueous alcohol. (A) Phenylhydrazine '(B)Semicarhanide (C) Hydroxylamine Semicarbazide Molar Molar Molar concn. of concn: of concn. of Concentration of semicarbazide 0.098, of carvone phenylsemihydroxyl0.018 mole per liter; solvent, 74% aqueous alcohol: ionic hydrazine ko X 102 carbazide ko X 10 amine ko X 101 strength, 0.398. 0.052 11.2 0.054 3.18 0.071 2.4 Formic acid F o r m a t e ion concentration concentration ko E. m. f. 9.5 ,069 ,074 2.63 .097 2.0 0.076 0.074 0.550 0.0350 ,092 9.1 .086 ,128 1.7 2.58 .114 ,111 ,618 .0338 Demonstrations of General Acid Catalysis.,152 ,148 ,691 ,0329 These runs were made to show that the reaction Hydroxylamine of the ketone with each carbonyl reagent is subConcentration of hydroxylamine 0.097, of carvone q.OY6 ject to general acid catalysis. The method used mole per liter; solvent, 485% aqueous alcohol, ionic was that of Brbhsted and Guggenheims in which strength, 0.225.
the same pH is maintained in each of several solutions containing different amounts of an acid HA. This is accomplished by adding to each solution enough of the anion of the acid, A-, to 18) Bronsted and Cugpenheim,
THIS JOI-RNAL, 49, 2 5 5 4 11927)
Hydrogen maleate concentration
0.032 .e48 ,064
THE INCREASE O F
Maleate ion concentration
ka
E. m. 1.
0.032
0.183
,048 064
206: ,233
0,379 368
I
.360
KINETICS AND MECHANISMS OF d-CARVONE
July, 1944
was not possible to convert these readings to actual PH values because of the unknown liquid junction potentials in the alcohol-water solutions. It can be seen that the ko’s increase with increasing acid concentration although the pH remains constant. This shows that these reactions are general acid catalyzed.
Discussion These experiments indicate that there can be little doubt that the reactions of d-carvone with phenylhydrazine, semicarbazide and hydroxylamine proceed by the same mechanism. The experimental facts which must be satisfied by a proposed mechanism may be summarized as follows: the reaction rates are proportional to the concentration of the ketone and to the concentration of the carbonyl reagent; the reactions are general acid catalyzed; and none of the reactions show any kinetic salt effect, which can be interpreted only as showing that the rate controlling step cannot be one between ions, but may be one between uncharged molecules or between an ion and an uncharged molecule. Accordingly, the velocity expression common to all three of these reactions, and which must therefore be the velocity expression for the rate controlling step, appears to be u = [RzCO][BNHzIkiL:[HAi]
as has been shown by Conant and Bartlett8 for the semicarbazide reaction. Since in our experiments the catalyzing acid was the conjugate acid, BNHs+, of the carbonyl reagent, this velocity expression becomes PI
= ki[RzCOl[BNHzI[BNHa+]
Our experimental data are in agreement with the following mechanism, which is a slightly modified version of that proposed by Hammett.4 It seems probable that the function of the acid catalyst is to increase the positive charge density on the carbonyl carbon. Hence, the first step in the reaction is (1) R E : : O
+ BNHa+
R&:O:H+
+ BNHz
which facilitates the acceptance of the free electron pair from the nitrogen of the carbonyl reagent
WITH CARBONYL
H+ :0: (2) R : C R
+
REAGENTS
1161
H+ :0: H R : C : N:B R H
H :N:B H
The remaining steps are
+ BNH2 J_ RzC(0H)NHB -+ BNHa+ (4) R2C(OH)KHB + BNH3+ )c R&(OHz)NHB+ + BNHz (5) RzC(OH)zNHB+ )c RzC=NHB+ + Hz0 (6) RzC=NHB+ + BNHz I _ RzC=NB + BNHa+ (3) RzC(OH)NH*B+
A mathematical analysis of these reactions shows that if the first step is rate controlling, the velocity expression for the reaction would be (1)
VI
= ki [ R E O ] [BNHa+]
If, instead, one of the other steps is rate controlling, the velocity expression would be
where K5 includes the relatively large constant concentration of water.
It follows that steps 1 , 2 and 5 cannot be rate controlling since they do not yield the necessary velocity expression. There are no data to decide between steps 3, 4 and 6, all of which have the proper velocity expression.
Summary 1. A comparative polarimetric study of the reactions between hydroxylamine, phenylhydrazine and semicarbazide with d-carvone shows that these reactions are general acid catalyzed, show no salt effect, and that their velocities are proportional to the concentrations of the ketone and carbonyl reagent. 2. These results indicate that these three reactions proceed by the same mechanism. 3. The data here presented are in agreement with Hammett’s mechanism for the semicarbazide reaction. SCHENLEY PARK PITTSBURGH, PENNA.
RECEIVED MARCH 6, 1944
